TrendsAnimal decisions are the result of multi-ple interacting cognitive phenotypes.

Proximate and ultimate explanations ofdecisions require the characterizationof cognitive phenotypes.

We provide an ontology of cognitivephenotypes in a JDM framework.

Empirical examples of successfullycharacterized cognitive phenotypesare highlighted.

1Department of Biological Sciences,University of Maryland BaltimoreCounty, Baltimore, MD 21250, USA2Department of Biology, DukeUniversity, Durham, NC 27705, USA3Department of Psychology, HunterCollege and the Graduate Center, CityUniversity of New York, New York, NY10065, USA4Department of Philosophy andReligious Studies, Louisiana StateUniversity, Baton Rouge, LA 70803,USA5Behavioral and Molecular EcologyGroup, Department of BiologicalSciences, University of Wisconsin-Milwaukee, Milwaukee, WI 53201,USA6Department of Ecology andEvolutionary Biology, University ofColorado Boulder, Boulder, CO 80309,USA7Santa Fe Institute, Santa Fe, NM87501, USA8Department of Psychology andCenter for Brain, Biology andBehavior, University of Nebraska-Lincoln, Lincoln, NE 68588, USA

*Correspondence: tamram@umbc.edu(T.C. Mendelson).@Twitter: @tamram

OpinionCognitive Phenotypes and theEvolution of Animal DecisionsTamra C. Mendelson,1,*,@ Courtney L. Fitzpatrick,2 MarkE. Hauber,3 Charles H. Pence,4 Rafael L. Rodríguez,5

Rebecca J. Safran,6 Caitlin A. Stern,7 and Jeffrey R. Stevens8

Despite the clear fitness consequences of animal decisions, the science ofanimal decision making in evolutionary biology is underdeveloped comparedwith decision science in human psychology. Specifically, the field lacks aconceptual framework that defines and describes the relevant componentsof a decision, leading to imprecise language and concepts. The ‘judgmentand decision-making’ (JDM) framework in human psychology is a powerful toolfor framing and understanding human decisions, and we apply it here tocomponents of animal decisions, which we refer to as ‘cognitive phenotypes’.We distinguish multiple cognitive phenotypes in the context of a JDM frame-work and highlight empirical approaches to characterize them as evolvabletraits.

Opening the Black Box of Animal Decision MakingThe fitness consequences of animal decisions (see Glossary), from mating to feeding to findingshelter, are relatively easy to characterize compared with the cognitive processes that drivethem. As for any complex trait, decisions emerge from the interaction of multiple components.We consider each of these components to be a unique cognitive phenotype (Figure 1). Byadopting this nomenclature, we emphasize that any component of the decision-making processcan vary among individuals, have genetic and environmental components of variation, and besubject to selection. Characterizing the cognitive phenotypes that underlie animal decisions isessential for understanding how decision making evolves.

For decades, researchers have struggled to describe the internal (cognitive) components ofanimal decisions. During the early 20th century, behaviorists such as B.F. Skinner approachedcognition as a black box, insisting that internal processes can never be inferred from obser-vations of behavior [1]. However, modern ethologists and behavioral ecologists have made astrong case for inferring cognition from carefully designed behavioral experiments [2–5]. More-over, as advancing technology reveals the neural architecture of cognition with unprecedentedprecision [6–9], we are poised to make substantive links between animal decisions in the wild,the cognitive phenotypes that produce those decisions, and the fitness consequences thatshape them [10,11].

Despite this progress, behavioral ecology lacks a cohesive framework for understanding thedecision-making process [12,13], often leading to imprecise language and concepts whendescribing animal decisions in natural systems. Researchers often use the same term to refer todifferent cognitive phenotypes. For example, ‘discrimination’ might be used to refer to theperceptual ability of an animal to distinguish two similar stimuli, or it might refer to the tendency tomate with conspecifics over heterospecifics; yet, the two situations involve different cognitiveand physiological processes (e.g., [7]). The absence of a cohesive decision framework is also

850 Trends in Ecology & Evolution, November 2016, Vol. 31, No. 11 http://dx.doi.org/10.1016/j.tree.2016.08.008

© 2016 Elsevier Ltd. All rights reserved.

GlossaryAssessment: extracting ameasurable value of a stimulus thatindicates quality or quantity.Bayes’ Rule: an algorithm forrationally updating a prior belief aboutthe state of the world given newevidence.Categorization: assigning two ormore similar stimuli to a set anddistinguishing between stimuli indifferent sets.Choice: the cognitive process ofselecting an action in the face ofalternatives.Cognition: the mechanisms bywhich animals acquire, process,store, and act on information [46].Cognitive phenotype: a discretecognitive feature that can bequantitatively measured and specifiedin terms of its neurobiology,physiology, or behavior (adaptedfrom [47]).Decision: an umbrella term thatrefers to the cognitive processes thatevaluate and select options to arriveat a course of action.Discrimination: distinguishing two ormore distinct stimuli.Judgment: making an inference ordrawing a conclusion about the stateof the world based on informationfrom the internal and externalenvironment.Ontology: a formal specification ofthe concepts or entities, and therelations among those entities,comprising a particular domain ofdiscourse.Preference: the cognitive encodingof a ranking of options.Recognition: responding predictablyto a previously experienced andremembered stimulus.

Informa�onSensa�onMemory

JudgmentDiscrimina�onCategoriza�onAssessmentRecogni�on

DecisionPreferenceChoice

Ac�onBehavior

Figure 1. Cognitive Phenotypes Underlying Animal Decisions. Animal behavior is initiated by information andculminates in action. The two center boxes represent the decision-making component of the process, which is modeled inpsychology as a combination of judgments and decisions. Terms commonly used in behavioral ecology are mapped ontothis framework and represent cognitive phenotypes that might vary among individuals and affect relative fitness. Feedbackis assumed within and between all components, and components and their interactions can be affected by context, that is,the dynamic social and ecological environment in which a cognitive phenotype is expressed.

problematic when behaviors are used to infer particular cognitive phenotypes, because differentcognitive processes can lead to the same behavioral output [1,12,14]. For example, a femalemight mate with a heterospecific because of any number of internal processes, including aninability to perceive a difference between conspecifics and heterospecifics or the absence of apreference (see below).

We propose that problems linking behavioral and cognitive phenotypes can be mitigated byadopting the robust conceptual framework that guides research in human decision making. Thisframework distinguishes two main categories of cognitive phenotypes: ‘judgments’ and‘decisions’. ‘Judgments’ refer to how individuals acquire and process information to arriveat an understanding of the situation or state of the world. ‘Decisions’ refer to the processes bywhich individuals use judgments to arrive at a course of action [15]. In other words, individualscan vary in two main ways: how they ‘see’ the world, and what they decide to do about it. Thisdistinction forms the foundation of human decision-making science and frames a highlyproductive research agenda in human psychology [12–16].

Here, we integrate this JDM framework into behavioral ecology, assigning cognitive phenotypessuch as discrimination, recognition, preference, and choice, to one of these two categories(Figure 1). This framework is a novel way for behavioral ecologists to consider animal decisionsand brings multiple advantages. First, it specifies two broad classes of process with fundamentallydifferent functions that can be modeled distinctly (e.g., Bayesian inference for judgments versusutility maximization for decisions; see below). Second, this framework emphasizes the cognitiveprocess over the behavioral end product. These cognitive phenotypes represent the relevant axesof variation, continuous or categorical, along which individuals can differ and upon which selectioncan act. Linking this variation with variation in fitness serves a fundamental goal of behavioralecology [17–19]. Third, identifying discrete cognitive phenotypes and anchoring them to a JDMframework clarifies their role in the decision-making process, which in turn allows us to identify andcompare homologous processes across taxa [20] and determine which experimental approachesreveal the phenotype of interest (Boxes 1 and 2). Finally, the study of decision making is amultidisciplinary science. Integrating the JDM framework into behavioral ecology facilitates ashared understanding of concepts and techniques across the fields of human and nonhumandecision making, advancing our understanding of decision making across disciplines.

Here, we begin by describing judgment and decision making as implemented in the field ofhuman psychology and discuss cases where it has been implicitly applied in behavioral ecology.We then identify concepts that form the core of a shared ontology, that is, a formal specification

Trends in Ecology & Evolution, November 2016, Vol. 31, No. 11 851

Box 1. Experimental Approaches

Behavioral ecologists typically study decision making using choice tests, in which they present options to an individualthat indicates choice with an action, such as moving toward an option or pulling a lever [48]. However, that action canresult from multiple judgments and decisions, and a choice test cannot always distinguish among the various cognitivephenotypes that might underlie the action. If an individual fails to choose one option over another, it is unclear whether theindividual failed to make the judgment (e.g. failed to discriminate) or decided not to act (e.g. lacked a preference for onestimulus over the other).

Therefore, additional information that complements a behavioral choice test is sometimes needed to specify the cognitivephenotypes underlying animal decisions. In particular, neural responses (e.g., electrophysiological recordings ormagnetic resonance imaging), physiological responses (e.g., breathing rate, heart rate, or hormone levels), or behavioralresponses other than choice (e.g., eye gaze or habituation–dishabituation) can be used to test for and characterize any ofthe cognitive phenotypes (Table I).

Discrimination can be revealed by presenting two stimuli simultaneously or sequentially and measuring neural, physio-logical, or behavioral responses. Different responses across stimulus types demonstrate discrimination.

To demonstrate recognition, neural, physiological, or behavioral variables should respond predictably to stimuli that havebeen previously experienced and differently to stimuli that have not been experienced. Recognition tests are distin-guished from simple discrimination tests by determining whether subjects distinguish specifically between familiar(remembered) and novel stimuli.

For categorization, measured variables should respond both similarly to multiple exemplars of a given set of stimuli anddifferently to multiple exemplars not in the set.

Demonstrating assessment requires comparing the responses of measured variables with features that are known toindicate the quantity or quality of a stimulus (e.g., ornaments or odors) to determine whether and how individuals attend tothose features.

Preference can be revealed by activity that exhibits a ranked response to varying stimuli.

Choice tests easily demonstrate choices by requiring an individual to choose between two or more simultaneouslypresented options. The neural and physiological correlates of that choice can be identified in a successful choice test andthen measured in a subsequent test to demonstrate choice even if action is impeded.

Table I. Experimental Approaches to Identifying Cognitive Phenotypesa

ResponseVariable

Techniques CognitivePhenotype

Refs

Neural Circuit visualization: in vivo imaging of active neural circuits Discrimination [7]

Electrophysiology: measures firing patterns of individual neurons Preference [8]

Magnetic resonance imaging: measures blood flow and neuralactivation of brain regions

Choice [49]

Physiological Breathing rate: measures the rate of respiration Discrimination [50]

Heart rate: measures the rate of heart activity Discrimination [51]

Hormone level: measures levels of hormones in blood, feces,saliva, or other bodily fluids

Preference [52]

Behavioral Eye gaze: measures the direction and duration of eye gaze Assessment [53]

Habitation–dishabituation: measures when individuals can detecta difference between stimuli to which they have versus have notbeen repeatedly exposed.

Categorization [54]

Behavioral trials: measures response to options using behavioralvariables other than direct selection of options or association time(e.g., type of display or intensity of activity)

Discrimination,categorization

[55]

aA host of techniques that measure neural, physiological, and behavioral responses can complement a behavioral choicetest to pinpoint which components of the decision-making process contribute to a given behavioral outcome. Examplesof cognitive phenotypes identified in the literature are provided, with references.

852 Trends in Ecology & Evolution, November 2016, Vol. 31, No. 11

of the concepts or entities that define an intellectual discipline and the relations among them [21].This ontology provides more than a shared lexicon; it generates fundamental conceptualdistinctions and can elucidate relations among cognitive processes that are frequently conflatedin the literature. We explore common usage of these terms within behavioral ecology, providedescriptions consistent with a JDM framework, and offer alternative experimental approachesthat could help identify each phenotype (Box 1), highlighting empirical examples (Box 2). Ourgoal is to advance a framework that facilitates an evolutionary perspective on animal decisions. AJDM view of decision making enables researchers to characterize relevant cognitive pheno-types, quantify variation in those phenotypes among individuals, and link fitness outcomes tothat variation.

The Judgment–Decision-Making DichotomySince the 1940s, the field of judgment and decision making has emerged from the integration ofcognitive psychology, social psychology, and economics to focus on two areas. One areaapplies principles of psychophysics and mathematical psychology to investigate how humansmake judgments about the state of the world; the other draws from economics to explore howhumans make decisions [15]. Although these areas developed through different approachesand questions, they represent an important conceptual distinction between making inferencesand making choices [12].

‘Judgments’ refer to how individuals assess information to arrive at an understanding of theirworld [15]. All organisms use various imperfect information (sensation, memory, etc.) to make ajudgment or inference about the true state of the world or to predict future states of the world.The primary measure of judgments is accuracy, that is, the correspondence between theperceived and true state of the world. Often Bayes’ rule is used to assess accuracy by

Box 2. Empirical Examples of Cognitive Phenotypes

Behavior Demonstrates Discrimination and Categorization without Preference in Sticklebacks

Kozak et al. [55] showed that male sticklebacks (Figure IA) court conspecific and heterospecific females with equivalentintensity; however, they vary the type of courtship display depending on the species, using behavior that is typical of thespecies being courted. This indicates an ability to discriminate the two species without a preference for one over theother. This experiment also demonstrated categorization; males exhibited the same display to multiple individuals of oneset and a different display to multiple individuals of another set.

Neuronal Spike Rates Demonstrate Categorization in Zebra Finches

Hauber et al. [9] measured the spike rates of single neurons in the auditory forebrain of female zebra finches (Figure IB) inresponse to male songs. Higher spike rates were recorded in response to conspecific songs compared with songs of twoheterospecifics; each stimulus category was represented by multiple individual exemplars. However, spike rates inresponse to the two heterospecifics did not differ, suggesting the categorization of conspecific song.

Eye Tracking Demonstrates Assessment In Peahens

Female peahens (Figure IC) have been shown to choose males based on the number of eyespots in their train and thelength of their long ‘fishtail’ feather. Yorzinski et al. [53] outfitted peahens with an eye-tracking device to monitor whichfeatures of a male were attended to by females during mate choice. Females paid most attention to the eyespots on thelower segments of the train as well as the fishtail, thus empirically supporting and quantifying assessment of thesefeatures.

Neuronal Spike Rates Demonstrate Preference in Gryllus bimaculatus

Kastarakos and Hedwig [8] showed that the strength of the phonotactic response of female Gryllus bimaculatus cricketsto variation in male chirp pulse duration varies continuously (Figure ID). When exposed to the same set of stimuli, activitypatterns of one neuron matched the graded female phonotaxis behavior, demonstrating a cognitively encoded ranking ofoptions (preference) that corresponded to female decisions.

Trends in Ecology & Evolution, November 2016, Vol. 31, No. 11 853

(B)

(C) (D)

(A)

Figure I. (A) Threespine stickleback Gasterosteus aculeatus (credit Piet Spaans), (B) Zebra finch Taeniopygia guttata(credit Keith Gerstung), (C) Indian peafowl Pavo cristatus (credit Jessica Yorzinksi), (D) Two-spotted cricket Gryllusbimaculatus (credit Gerd Rossen/www.digital-nature-photography.com).

determining whether an individual is accurately updating its belief about the state of the worldgiven new information [22].

By contrast, decisions refer to how individuals determine a course of action or choose amongdifferent options [15]. Whereas judgments are typically modeled in a Bayesian framework, muchof the study of decisions has come from economics and utility theory: to continue performing wellin the market, an individual should choose options that offer the highest utility [23]. There is nomeasure of accuracy for decisions because they do not directly refer to the state of the world.The question of what is a good decision depends upon how well that decision coheres to somecriterion. In economics, that criterion is expected utility maximization; in evolutionary biology, it isfitness maximization.

Although conceptually and empirically separable, judgments and decisions are nonethelessinextricably linked. Judgments are rarely made without a subsequent selection among options;decisions often require some evaluation of the state of the world; and the processes can feedback on one another [12]. For example, when a female spider encounters a male conspecific,she may categorize him as a potential mate (judgment #1) and simultaneously or sequentiallyassess his quality (judgment #2). Depending on these judgments, she may allow him toapproach (decision #1) and then mate with him (decision #2). Early decisions can modifysubsequent judgments, which in turn can modify subsequent decisions. For example, if shedecides to eat the male instead of mating with him, assessment may focus on nutritional quality.Thus, as for any decision, copulation requires both judgments and decisions that can actsequentially, simultaneously, and iteratively.

We are not the first to apply this kind of framework to behavioral ecology. Blumstein and Bouskila[14] presented an ‘assessment’ and decision-making framework, where assessment was

854 Trends in Ecology & Evolution, November 2016, Vol. 31, No. 11

essentially equivalent to psychology's judgment. They pointed out that cognitive processes aredifficult to glean from behavior alone, and that knowledge of ‘informational states’ (judgments) iscritical (see Figure 1 in [14]). We extend the ideas of Blumstein and Bouskila [14], but rather thaninventing a new framework and associated terminology, we suggest that JDM, as a productivepsychological framework already in place, can serve to clarify thinking in behavioral ecology andhelp identify the proximate bases and ultimate consequences of animal decisions.

Developing a Decision-Making OntologyBlumstein and Bouskila [14] demonstrated that behavioral ecologists might already implicitlydistinguish between judgments and decisions. Nonetheless, many terms that describe thecomponents of animal decisions are used inconsistently, interchangeably, and without a guidingframework in which to organize and distinguish them. To clarify terms in a way that is useful forbehavioral researchers, we explored the use of six common and often interchangeably usedterms from behavioral ecology: assessment, categorization, discrimination, recognition, pref-erence, and choice. We conducted a targeted literature search to examine how these termswere defined or used in three well-studied behavioral domains: mating, foraging, and habitatselection. For each term, we searched in Web of Sciencei using the stem of the term (e.g.‘recogn*’ or ‘choice’), other terms that specified one of the three behavioral domains (e.g.,‘forag*’, ‘mate’, or ‘habitat’), and the terms ‘definition’ or ‘defined’. For each term, we identified aminimum of two articles per behavioral domain with clearly labeled definitions, as well as multipleadditional articles from which definitions could be gleaned from usage. Based on this qualitativeanalysis, here we summarize usages across domains for each term and offer new or modifieddescriptions of each term that distinguish them one from another within the JDM framework.

JudgmentThe concept of judgment is not widely applied in behavioral ecology. We advocate a descriptionconsistent with the psychological literature, as ‘making an inference or drawing a conclusionabout the state of the world based on information from the internal and external environment’.Four processes associated with judgments (discrimination, categorization, assessment, andrecognition) are commonly used in behavioral ecology.

DiscriminationDiscrimination is arguably the most basic cognitive process involved in a judgment. Environ-ments (stimuli) vary along multiple dimensions, and distinguishing natural variants can be criticalto the fitness of an individual. The concept of discrimination is used broadly in all three domains ofbehavioral ecology examined here. In some cases, it is defined as a cognitive process precedingaction [24], but is more commonly defined as an action, when animals behave differently towardone stimulus versus another [25–27]. We consider discrimination, rather than an action, to be thecognitive process of ‘distinguishing two or more distinct stimuli’.

CategorizationCategorization is another process used to infer the state of the world and is predicated ondiscrimination. Categorization is used often in the mating and foraging domains of behavioralecology, but rarely in the context of habitat selection. Unlike discrimination, categorization ismost commonly described as a cognitive process [28,29], although actions can be used toidentify it. We describe categorization as ‘assigning two or more similar stimuli to a set anddistinguishing between stimuli in different sets’. Therefore, demonstrating categorizationrequires multiple test stimuli that either do or do not belong to a set of interest.

AssessmentAssessment is used across the three examined domains of behavioral ecology. Most commonly,assessment describes the acquisition of information related to the quantity or quality of a given

Trends in Ecology & Evolution, November 2016, Vol. 31, No. 11 855

stimulus, implicitly or explicitly referring to the fitness consequences of that stimulus [30,31].Unlike discrimination or categorization, the concept of assessment applies to a single stimulusrather than two or more and is best described as ‘extracting a measurable value of a stimulusthat indicates quality or quantity’.

RecognitionRecognition is frequently used in the mating domain of behavioral ecology, but less so in foragingor habitat selection. In the mating literature, recognition describes a response to any member ofa certain category, as in ‘mate recognition’ or ‘kin recognition’ (e.g., [32,33]). By contrast, inpsychology, the concept of recognition generally emphasizes a role of memory [34,35]. To ‘re-cognize’ a stimulus is to experience and process it again, meaning that the stimulus has beenpreviously processed. Therefore, recognition is ‘responding predictably to a previously experi-enced and remembered stimulus’. A good example of recognition is individual recognition inpaper wasps, in which individuals are remembered based on facial patterns [36]; however, manycases of mate recognition [37], competitor recognition [38], kin recognition, predator recogni-tion, and species recognition, in which animals respond to sets of individuals rather thanremembered stimuli, are better described as categorization than recognition. Framing thesephenomena as categorization allows us to apply existing theories in psychology to behavioralecology (e.g., exemplar versus prototype theories of categorization [39]) to better understandhow these judgments are made.

Decision‘Decision’ is an umbrella term that refers to the cognitive processes that evaluate and selectoptions to arrive at a course of action. A decision itself is not an overt action or behavior;individuals may decide before acting. Yet, because decisions are further ‘downstream’ in thedecision-making process (Figure 1), behavioral tests may provide more direct measures ofdecisions than of judgments.

PreferencePreference is used in all three behavioral ecology domains examined and typically is described asthe act of selecting one or some options over others (e.g., [25,40,41]). Certainly behavioralexperiments can identify preferences, but we distinguish an internal process of preference fromits behavioral outcome. For example, neurons in the anterior protocerebrum of female Gryllusbimaculatus crickets fire at different rates depending on the pulse duration of male calls, withsome durations producing higher firing rates than others [8] (Box 2). Therefore, this neural‘ranking’ of stimuli is represented internally and independently of its behavioral outcome. Wedescribe preference as ‘the cognitive encoding of a ranking of options’.

ChoiceChoice is used ubiquitously in behavioral ecology. Perhaps more than any other component of adecision-making process, choice is defined as an action, such as a behavior that restricts thepotential set of mates [42], a specific type of foraging behavior that is driven by preference [43], orthe probability of settling in a particular habitat [44]. However, although these actions result froma choice, an organism arrives at that choice before the motor activity typically measured in anexperiment. For example, an individual may choose a mate but be prevented from mating due tothe presence of a predator. Therefore, we describe choice as ‘the cognitive process of selectingan action in the face of alternatives’.

Although no one example can illustrate all of these cognitive phenotypes perfectly, imagine thefollowing case. A sample of female crickets faces a choice between a conspecific and hetero-specific call in a phonotaxis trial (Figure 2), and their behavior varies. Some females approach theconspecific call, whereas others spend an equal amount of time with each call type. Multiple

856 Trends in Ecology & Evolution, November 2016, Vol. 31, No. 11

Figure 2. A Typical Phonotaxis Trial Offering A Female Cricket A Choice between Two Song Types.

Outstanding QuestionsHow do cognitive phenotypes varyamong individuals, genotypes, envi-ronments, etc.?

How does each cognitive phenotypeinfluence and co-vary with others?

What are the proximate bases of cog-nitive phenotypes (i.e., mapping geno-type to phenotype)?

To what extent does variation in cogni-tive phenotypes influence fitness?

What additional cognitive phenotypescontribute to animal decisions?

hypotheses, targeting different cognitive phenotypes, could explain the difference in behavior,and Box 1 offers guidance on testing each of these hypotheses: (i) some females are able todiscriminate between the two call types (i.e., perceive them as different), whereas other femalescannot; (ii) some females recognize one call type from prior experience and others fail toremember it; (iii) some females categorize conspecific and heterospecific calls into differentsets, whereas others categorize them together; and (iv) some females are able to assess therelevant features of the calls (e.g., pulse rate), whereas others cannot detect these features.

Those four hypotheses explain behavioral variation as a result of differences in judgment;however, females also can differ in decision phenotypes: (i) some females might have apreference for trait values that correspond with conspecific calls, whereas others rank conspe-cific and heterospecific calls equally; and (ii) all females might prefer conspecific calls, but someare motivated to make a choice, while others are not, perhaps because they recently mated.

Thus, clarifying the concepts that we use to describe and quantify animal decisions facilitatesdiscussion and hypothesis testing within and across disciplines. Behavioral ecologists mightalready implicitly acknowledge the broader distinction of the JDM framework; however, byexplicitly speaking a common language based on this framework, behavioral ecologists canleverage existing models (e.g., Bayesian inference or optimality) and experimental designs (e.g.,habituation–dishabituation or eye tracking) from psychology to define and characterize cognitivephenotypes, test hypotheses about behavioral variation, and predict evolutionary outcomes innatural systems.

Concluding Remarks and Future DirectionsWith carefully designed studies that distinguish cognitive phenotypes, we can identify thegenetic and neurophysiological bases of individual variation and achieve the central aim ofbehavioral ecology: linking the causes of animal decisions with their fitness consequences.However, whereas the fitness consequences of a decision may be relatively straightforward, itscauses have been concealed inside the black box that connects information with action. TheJDM framework begins to open that black box and provides an ontology of cognitive pheno-types that researchers can use to identify homologous phenotypes and characterize individualvariation. However, as implemented in human psychology, the JDM framework lacks an explicitevolutionary context. Behavioral ecology provides this evolutionary perspective and can offerpredictions about which judgments are likely to be subject to selection (i.e., those that affect

Trends in Ecology & Evolution, November 2016, Vol. 31, No. 11 857

fitness) and, therefore, for which we expect to observe accurate judgments [12]. Behavioralecology also offers a clear currency of maximization for decisions (i.e., relative fitness) [45].

The core list of cognitive phenotypes discussed here presents behavioral ecologists with multipleresearch directions (see Outstanding Questions). As researchers with access to the broadestrange of animal systems and decision-making contexts, behavioral ecologists are well posi-tioned to identify the neural, physiological, and genetic bases of cognitive phenotypes, thusadvancing our understanding of how animal decisions are generated and how decision makingevolves.

AcknowledgmentsThis work was supported by the National Evolutionary Synthesis Center (NESCent), NSF #EF-0905606. The authors wish to

thank NESCent staff for their support and additional members of the working group ‘Toward a unified evolutionary theory of

decision making in animals’ for helpful discussions.

Resourcesi http://wokinfo.com/

References

1. Skinner, B.F. (1938) The Behavior of Organisms, Appleton-

Century

2. Dukas, R. (1998) Cognitive Ecology: The Evolutionary Ecology ofInformation Processing and Decision Making (Vol. 1), University ofChicago Press

3. Dukas, R. and Ratcliffe, J.M. (2009) Cognitive Ecology II, Universityof Chicago Press

4. Gould, J.L. (2004) Thinking about thinking: how Donald R. Griffin(1915-2003) remade animal behavior. Anim. Cogn. 7, 1–4

5. Griffin, D.R. and Speck, G.B. (2004) New evidence of animalconsciousness. Anim. Cogn. 7, 5–18

6. Russ, B.E. et al. (2007) Neural and behavioral correlates of auditorycategorization. Hearing Res. 229, 204–212

7. Clowney, E.J. et al. (2015) Multimodal chemosensory circuitscontrolling male courtship in drosophila. Neuron 87, 1036–1049

8. Kostarakos, K. and Hedwig, B. (2012) Calling song recognition infemale crickets: temporal tuning of identified brain neuronsmatches behavior. J. Neurosci. 32, 9601–9612

9. Hauber, M.E. et al. (2013) Experience dependence of neuralresponses to different classes of male songs in the primary auditoryforebrain of female songbirds. Behav. Brain Res. 243, 184–190

10. Buchanan, K.L. et al. (2013) Condition dependence, developmen-tal plasticity, and cognition: implications for ecology and evolution.Trends Ecol. Evol. 28, 290–296

11. Thornton, A. et al. (2014) Toward wild psychometrics: linking indi-vidual cognitive differences to fitness. Behav. Ecol. 25, 1299–1301

12. Stevens, J.R. (2008) The evolutionary biology of decision making.In Better than Conscious? Decision Making, the Human Mind, andImplications for Institutions (Engel, C. and Singer, W., eds), pp.285–304, MIT Press

13. Hammerstein, P. and Stevens, J.R. (2012) Evolution and theMechanisms of Decision Making, MIT Press

14. Blumstein, D.T. and Bouskila, A. (1996) Assessment and decisionmaking in animals: a mechanistic model underlying behavioralflexibility can prevent ambiguity. Oikos 77, 569–576

15. Goldstein, W.M. and Hogarth, R.M. (1997) Judgment and decisionresearch: some historical context. In Research on Judgment andDecision Making: Currents, Connections, and Controversies(Goldstein, W.M. and Hogarth, R.M., eds), pp. 3–65, CambridgeUniversity Press

16. Goldstein, W.M. and Hogarth, R.M., eds (1997) Research onJudgment and Decision Making: Currents, Connections, andControversies,, Cambridge University Press

17. Dall, S.R.X. et al. (2012) Variation in decision making. In Evolutionand the Mechanisms of Decision Making (Hammerstein, P. andStevens, J.R., eds), pp. 243–272, MIT Press

858 Trends in Ecology & Evolution, November 2016, Vol. 31, No.

18. Rowe, C. and Healy, S.D. (2014) Measuring variation in cognition.Behav. Ecol. 25, 1287–1292

19. Morand-Ferron, J. and Quinn, J.L. (2015) The evolution of cogni-tion in natural populations. Trends Cog. Sci. 19, 235–237

20. Chittka, L. et al. (2012) What is comparable in comparative cogni-tion? Phil. Trans. Roy. Soc. B 367, 2677–2685

21. Hofweber, T. (2014) Logic and ontology. In The Stanford Ency-clopedia of Philosophy (Zalta, E.N., ed.), http://plato.stanford.edu/entries/logic-ontology/

22. Tenenbaum, J.B. et al. (2011) How to grow a mind: Statistics,structure, and abstraction. Science 331, 1279–1285

23. Varian, H.R. (2014) Intermediate Microeconomics: A ModernApproach. (9th edn), WW Norton & Company

24. Niklitschek, E.J. and Secor, D.H. (2010) Experimental and fieldevidence of behavioural habitat selection by juvenile Atlantic Aci-penser oxyrinchus oxyrinchus and shortnose Acipenser breviros-trum sturgeons. J Fish Biol. 77, 1293–1308

25. Campbell, D.L.M. and Hauber, M.E. (2009) Spatial and behav-ioural measures of social discrimination by captive male zebrafinches: implications of sexual and species differences for recog-nition research. Behav. Proc. 80, 90–98

26. Lackey, A.C. and Boughman, J.W. (2014) Female discriminationagainst heterospecific mates does not depend on mating habitat.Behav. Ecol. 25, 1256–1267

27. Herrera-Varela, M. et al. (2014) Habitat discrimination by gravidAnopheles gambiae sensu lato: a push–pull system. Malaria Jour-nal 13, 1

28. Menzel, R. and Giurfa, M. (2001) Cognitive architecture of a mini-brain: the honeybee. Trends Cog. Sci. 5, 62–71

29. Baugh, A.T. et al. (2008) Categorical perception of a natural,multivariate signal: mating call recognition in Túngara frogs. Proc.Natl. Acad. Sci. 105, 8985–8988

30. Seeley, T.D. (1989) Social foraging in honey bees: how nectarforagers assess their colony's nutritional status. Behav. Ecol.Sociobiol. 24, 181–199

31. Bonduriansky, R. (2001) The evolution of male mate choice ininsects: a synthesis of ideas and evidence. Biol. Rev. 76, 305–339

32. Ryan, M.J. and Rand, A.S. (1993) Species recognition and sexualselection as a unitary problem in animal communication. Evolution47, 647–657

33. Macedonia, J.M. et al. (2013) Species recognition of color andmotion signals in Anolis grahami: evidence from responses tolizard robots. Behav. Ecol. 24, 846–852

34. Santrock, J.W. (2005) Psychology, McGraw-Hill

35. Weiten, W. (2015) Psychology: Themes and Variations, CengageLearning

11

36. Tibbetts, E.A. and Dale, J. (2007) Individual recognition: it is goodto be different. Trends in Ecology & Evolution 22, 529–537

37. Reeve, H.K. (1989) The evolution of conspecific acceptancethresholds. Am. Nat. 33, 407–435

38. Grether, G.F. (2011) The neuroecology of competitor recognition.Int. Comp. Biol. 51, 807–818

39. Ashby, F.G. and Maddox, W.T. (2005) Human category learning.Ann. Rev. Psych. 56, 149–178

40. Krivan, V. (2010) Evolutionary stability of optimal foraging:partial preferences in the diet and patch models. J. Theor. Biol.267, 486–494

41. Hollander, F.A. et al. (2011) Maladaptive habitat selection of amigratory passerine bird in a human-modified landscape. PLoSONE 6, e25703

42. Wiley, R.H. and Poston, J. (1996) Perspective: indirect matechoice, competition for mates, and coevolution of the sexes.Evolution 50, 1371–1381

43. Jackson, R.R. et al. (2005) A spider that feeds indirectly onvertebrate blood by choosing female mosquitoes as prey. Proc.Natl. Acad. Sci. 102, 15155–15160

44. Ravigné, V. et al. (2009) Live where you thrive: joint evolution ofhabitat choice and local adaptation facilitates specialization andpromotes diversity. Am. Nat. 174, E141–E169

45. Kacelnik, A. (2006) Meanings of rationality. In Rational Animals?(Hurley, S. and Nudds, M., eds), pp. 87–106, Oxford UniversityPress

46. Shettleworth, S.J. (2010) Cognition, Evolution, and Behavior,Oxford University Press

47. Santangelo, S.L. and Jagaroo, V., eds (2016) Cognitive andBehavioral Phenotypes, Springer

48. Dougherty, L.R. and Shuker, D.M. (2015) The effect of experimen-tal design on the measurement of mate choice: a meta-analysis.Behav. Ecol. 26, 311–319

49. Voss, H.U. et al. (2010) Altered auditory BOLD response to con-specific birdsong in zebra finches with stuttered syllables. PLoSONE 5, e14415

50. Barreto, R.E. et al. (2003) Ventilatory frequency indicates visualrecognition of an allopatric predator in naıve Nile tilapia. Behav.Proc. 60, 235–239

51. Hauber, M.E. et al. (2002) Discrimination between host songs bybrood parasitic brown-headed cowbirds (Molothrus ater). Anim.Cogn. 5, 129–137

52. Tchernichovski, O. et al. (1998) Context determines the sex appealof male zebra finch song. Anim. Behav. 55, 1003–1010

53. Yorzinski, J.L. et al. (2013) Through their eyes: selective attentionin peahens during courtship. J. Exp. Biol. 216, 3035–3046

54. Wyttenbach, R.A. et al. (1996) Categorical perception of soundfrequency by crickets. Science 273, 1542–1544

55. Kozak, G.M. et al. (2009) Sex differences in mate recognition andconspecific preference in species with mutual mate choice. Evo-lution 63, 353–365

Trends in Ecology & Evolution, November 2016, Vol. 31, No. 11 859